<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Ai-Safety on Martin's Blog</title><link>https://mac.sploit.dk/tags/ai-safety/</link><description>Recent content in Ai-Safety on Martin's Blog</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Mon, 14 Sep 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://mac.sploit.dk/tags/ai-safety/index.xml" rel="self" type="application/rss+xml"/><item><title>The Mythos 5 Incidents: What Anthropic and AISI Documented</title><link>https://mac.sploit.dk/blog/mythos-5-incidents-anthropic-aisi/</link><pubDate>Mon, 14 Sep 2026 00:00:00 +0000</pubDate><guid>https://mac.sploit.dk/blog/mythos-5-incidents-anthropic-aisi/</guid><description>&lt;p&gt;Anthropic and the UK AI Security Institute have now published separate accounts of Claude Mythos 5 taking cyber evaluations onto the real internet. I wanted to know whether they describe the same failure pattern, and what the released transcript actually supports.&lt;/p&gt;&#10;&lt;p&gt;The short answer is yes, with an important qualification: these were different evaluations. Anthropic&amp;rsquo;s transcript covers a PyPI incident on 18 July 2026. AISI&amp;rsquo;s “Doing Life” runs took place a week later.&lt;/p&gt;</description></item></channel></rss>